Impact of protein conformational diversity on AlphaFold predictions

Abstract

Motivation After the outstanding breakthrough of AlphaFold in predicting protein 3D models, new questions appeared and remain unanswered. The ensemble nature of proteins, for example, challenges the structural prediction methods because the models should represent a set of conformers instead of single structures. The evolutionary and structural features captured by effective deep learning techniques may unveil the information to generate several diverse conformations from a single sequence. Here we address the performance of AlphaFold2 predictions obtained through ColabFold under this ensemble paradigm. Results Using a curated collection of apo-holo pairs of conformers, we found that AlphaFold2 predicts the holo form of a protein in ∼70% of the cases, being unable to reproduce the observed conformational diversity with the same error for both conformers. More importantly, we found that AlphaFold2's performance worsens with the increasing conformational diversity of the studied protein. This impairment is related to the heterogeneity in the degree of conformational diversity found between different members of the homologous family of the protein under study. Finally, we found that main-chain flexibility associated with apo-holo pairs of conformers negatively correlates with the predicted local model quality score plDDT, indicating that plDDT values in a single 3D model could be used to infer local conformational changes linked to ligand binding transitions. Availability Data and code used in this manuscript are publicly available at https://gitlab.com/sbgunq/publications/af2confdiv-oct2021 Supplementary Information Supplementary data is available at the journal's web site.

Structural insights into the role of the WW2 domain on tandem WW–PPxY motif interactions of oxidoreductase WWOX

“…This is in line with previous reports that AlphaFold2 often generates bound conformations even in the absence of the ligand ( e.g. , ( 43 , 44 )). For similar reasons, in its quest for the most stable conformation, AlphaFold2 cannot be used to assess effects of point mutations, since it relates to point mutations as “local noise” and will by default converge on the same result ( e.g.…”

AlphaFold, Artificial Intelligence (AI), and Allostery

“…AlphaFold has overcome age-long bottlenecks and forcefully bared the power of artificial intelligence (AI) in biological research. − AlphaFold has combined numerous deep learning innovations to predict the three-dimensional (3D) structures of proteins at or near experimental scale resolution, inspiring the community (including us) to rethink studies of function, evolution, and disease (e.g., refs − ). The sheer volume of the rapidly generated accurate structures argues that new, ambitious, frontier-pushing studies will emerge.…”

Challenges in describing the conformation and dynamics of proteins with ambiguous behavior

“…These results show again that AlphaFold2 is excellent at defining a single low-energy state for a given protein sequence if it exists, but that the context of the protein and possible ambiguous behavior is more difficult to capture. Indeed, in relation to conformational diversity as observed in the PDB from apo-holo pairs of conformers for the same protein ( Saldaño et al, 2022 ), AlphaFold2 predicts the holo form in ∼70% of cases but is unable to capture both states. As the conformational diversity between the apo/holo states increases, its prediction performance also worsens.…”